More changes
Browse files- app.py +7 -3
- feature_extraction.py +8 -5
app.py
CHANGED
@@ -87,18 +87,22 @@ if uploaded_file is not None:
|
|
87 |
"Punk","Bebop", "Pop", "R&B", "Country", "Rap & Hip-Hop", "Soul"]
|
88 |
class_indices = {i: class_name for i, class_name in enumerate(class_names)}
|
89 |
|
90 |
-
features_list = audio_splitting.split_audio(uploaded_file)
|
91 |
features = feature_extraction.scale(features_list)
|
92 |
|
93 |
# st.write(features)
|
94 |
# Features Dataframe
|
95 |
df = pd.DataFrame({
|
96 |
-
"fname": ["
|
|
|
97 |
})
|
98 |
st.dataframe(
|
99 |
df,
|
100 |
column_config={
|
101 |
-
"name": "Features"
|
|
|
|
|
|
|
102 |
}
|
103 |
)
|
104 |
|
|
|
87 |
"Punk","Bebop", "Pop", "R&B", "Country", "Rap & Hip-Hop", "Soul"]
|
88 |
class_indices = {i: class_name for i, class_name in enumerate(class_names)}
|
89 |
|
90 |
+
features_list,val_list = audio_splitting.split_audio(uploaded_file)
|
91 |
features = feature_extraction.scale(features_list)
|
92 |
|
93 |
# st.write(features)
|
94 |
# Features Dataframe
|
95 |
df = pd.DataFrame({
|
96 |
+
"fname": ["Chroma_STFT"],
|
97 |
+
"Values": val_list
|
98 |
})
|
99 |
st.dataframe(
|
100 |
df,
|
101 |
column_config={
|
102 |
+
"name": "Features",
|
103 |
+
"Values": st.column_config.LineChartColumn(
|
104 |
+
"Graph Values",y_min=0,y_max = 10000
|
105 |
+
)
|
106 |
}
|
107 |
)
|
108 |
|
feature_extraction.py
CHANGED
@@ -31,16 +31,19 @@ short_field = Fields[2:]
|
|
31 |
|
32 |
def all_feature_extraction(audio_path, sample_rate=22050):
|
33 |
data_list = []
|
|
|
|
|
34 |
audio_df, sr = librosa.load(audio_path, sr=22050)
|
35 |
-
print("\n",audio_df)
|
36 |
data_list.append(audio_path)
|
37 |
-
print(audio_path)
|
38 |
data_list.append(len(audio_df))
|
39 |
-
|
|
|
40 |
# 1. Chroma STFT
|
41 |
chroma_stft = librosa.feature.chroma_stft(y=audio_df, hop_length=512)
|
42 |
chroma_stft_mean = np.mean(chroma_stft)
|
43 |
chroma_stft_var = np.var(chroma_stft)
|
|
|
|
|
44 |
data_list.append(chroma_stft_mean)
|
45 |
data_list.append(chroma_stft_var)
|
46 |
|
@@ -103,8 +106,8 @@ def all_feature_extraction(audio_path, sample_rate=22050):
|
|
103 |
for mean, var in mfcc_list:
|
104 |
data_list.append(mean)
|
105 |
data_list.append(var)
|
106 |
-
|
107 |
-
return data_list
|
108 |
|
109 |
def scale(initial_features):
|
110 |
final_features = initial_features[2:]
|
|
|
31 |
|
32 |
def all_feature_extraction(audio_path, sample_rate=22050):
|
33 |
data_list = []
|
34 |
+
val_field = []
|
35 |
+
|
36 |
audio_df, sr = librosa.load(audio_path, sr=22050)
|
|
|
37 |
data_list.append(audio_path)
|
|
|
38 |
data_list.append(len(audio_df))
|
39 |
+
|
40 |
+
|
41 |
# 1. Chroma STFT
|
42 |
chroma_stft = librosa.feature.chroma_stft(y=audio_df, hop_length=512)
|
43 |
chroma_stft_mean = np.mean(chroma_stft)
|
44 |
chroma_stft_var = np.var(chroma_stft)
|
45 |
+
|
46 |
+
val_field.append(chroma_stft)
|
47 |
data_list.append(chroma_stft_mean)
|
48 |
data_list.append(chroma_stft_var)
|
49 |
|
|
|
106 |
for mean, var in mfcc_list:
|
107 |
data_list.append(mean)
|
108 |
data_list.append(var)
|
109 |
+
|
110 |
+
return [data_list,val_field]
|
111 |
|
112 |
def scale(initial_features):
|
113 |
final_features = initial_features[2:]
|